Qualités psychométriques de l’adaptation française du Family Relationship Index (FRI)

Abstract

The aim of the present study was to investigate the psychometric properties of the French adaptation of the Family Relationship Index (FRI) from Moos and Moos. The FRI is a self-report inventory which consists of 27 items assessing family relations. It is composed of three dimensions: family cohesion, family expressiveness (of feelings and emotions) and family conflict. The FRI was translated and adapted into French according to published recommendations. After appropriate cultural adaptations, the scale was administered to a sample of 976 students with a mean age of 21.9 years and 43.5% of men. The participants completed the FRI and three other questionnaires: the Family Adaptability and Cohesion Scale (FACES), the Satisfaction with Life Scale and the General Health Questionnaire. Confirmatory factor analyses were used to test different models with one and three factors. The psychometric properties of the short version of the FRI, proposed by Kissane and Bloch (2002) and composed of 12 items, were also studied. Confirmatory factor analyses showed that the three factors solution was more relevant that the one factor solution (for Khi(2)/ddl, Root mean square error of approximation [RMSEA], Root mean square residual [RMR], Goodness of fit index [GFI], Adjusted goodness of fit [AGFI] and Comparative fit index [CFI]). However, three items belonging to family expression explained a small variance. Therefore, a version consisting of 24 items seemed more appropriate than the 27 items version proposed by Moos and Moos. Cultural differences may explain these results. Internal consistency was satisfactory for cohesion (0.79) and conflict (0.71) but weak for expression (0.55 for 27 items version and 0.62 for 24 items version). One month test-retest reliability showed high correlations for the three dimensions (from 0.77 to 0.85). Correlation between the cohesion subscale of the FRI and the cohesion subscale of the FACES was high (0.77), showing a good convergent validity. The correlations between the three FRI dimensions and the Satisfaction with Life Scale and the General Health Questionnaire were quite low (from -0.31 to 0.41). High family cohesion and high family expression were associated with high life satisfaction and good mental health. Conversely, high family conflict was associated with low life satisfaction and weak mental health. These results support the criterion validity of the FRI. Concerning the 12 items version of the FRI, the factor analysis results showed very good psychometric qualities. However, this short version had lower internal consistency (which ranged between 0.50 and 0.71), test-retest reliability (which ranged 0.68 and 0.81), convergent (0.69 for cohesion) and criterion validity (from -0.21 to 0.37 for the Satisfaction with Life Scale and from -0.28 to 0.19 for the General Health Questionnaire) than the longer versions. The results of the present study show that the FRI's 24 items version seems to be the more relevant. Nevertheless, the 12 items version shows interesting qualities. Further studies should confirm these results on other samples. Given the lack of French-language surveys assessing family relations, the FRI will be a useful tool for research and clinical practice.